Human neutrophil lipocalin, procalcitonin, c-reactive protein, and leucocyte count for prediction of bacterial sepsis in emergency department patients.

BACKGROUND

Delayed identification of bacterial sepsis undermines the initiation of antibiotic and other time-sensitive treatments in the emergency department (ED). We aimed to investigate the performance of human-neutrophil lipocalin (HNL), procalcitonin (PCT), C-reactive protein (CRP), and leucocyte count in conjunction with clinical scores for the early prediction of bacterial sepsis.

METHODS

Patients presenting to the emergency department (ED) with a suspected infection and a national early warning score (NEWS) ≥ 2 at triage were screened for eligibility. The study biomarkers were measured at ED presentation. The primary outcome was bacterial sepsis, defined as an acute bacterial infection and an increase of ≥ 2 points in the sequential organ failure assessment (SOFA) score (Sepsis-3 criteria). The diagnostic accuracy of the biomarkers for bacterial sepsis was calculated using receiver operating curve (ROC) analysis and its area under the curve (AUC) with 95% confidence intervals (CI).

RESULTS

In total, we included 421 patients, of whom 155 (36.8%) had bacterial sepsis. For the prediction of bacterial sepsis, PCT outperformed the other biomarkers with an AUC (95% CI) of 0.77 (0.72-0.82), compared to HNL 0.72 (0.67-0.77), CRP 0.71 (0.66-0.76), and leucocyte count 0.64 (0.59-0.70). A combination of serum HNL with NEWS and SOFA at presentation had the best predictive value for bacterial sepsis at 48 h after ED presentation (AUC 0.83).

CONCLUSION

Our findings do not support the routine use of serum HNL as a tool in the prediction of bacterial sepsis in patients presenting to the ED. A combination of biomarkers (serum HNL or CRP plus leucocytes) with NEWS and SOFA at presentation outperformed inflammatory biomarkers used individually in the prediction of bacterial sepsis.